Calibrating wavelet neural networks by distance orientation similarity fuzzy C-means for approximation problems

Ong, Pauline and Zainuddin, Zarita (2016) Calibrating wavelet neural networks by distance orientation similarity fuzzy C-means for approximation problems. Applied Soft Computing, 42 . pp. 156-166. ISSN 15684946

[img]
Preview
PDF
1441Kb

Official URL: http://dx.doi.org/10.1016/j.asoc.2016.01.042

Abstract

Improperly tuned wavelet neural network (WNN) has been shown to exhibit unsatisfactory generaliza-tion performance. In this study, the tuning is done by an improved fuzzy C-means algorithm, that utilizesa novel similarity measure. This similarity measure takes the orientation as well as the distance intoaccount. The modified WNN was first applied to a benchmark problem. Performance assessments withother approaches were made subsequently. Next, the feasibility of the proposed WNN in forecasting thechaotic Mackey–Glass time series and a real world application problem, i.e., blood glucose level predic-tion, were studied. An assessment analysis demonstrated that this presented WNN was superior in termsof prediction accuracy.

Item Type:Article
Uncontrolled Keywords:clustering; distance similarity; function approximation; orientation similarity; time series; wavelet neural networks
Subjects:Q Science > QA Mathematics > QA297 Numerical analysis. Analysis
Divisions:Faculty of Mechanical and Manufacturing Engineering > Department of Engineering Mechanics
ID Code:8005
Deposited By:Normajihan Abd. Rahman
Deposited On:10 May 2016 12:38
Last Modified:10 May 2016 12:38

Repository Staff Only: item control page